Birth · Newborn
Neonatal Sepsis Calculator (EOS)
Early-onset sepsis (EOS) risk-based calculator — simplified educational implementation of the Kaiser Permanente / Kuzniewicz 2017 model. Reduces unnecessary antibiotic exposure by 60-70 % vs categorical risk-factor algorithms.
Last reviewed 25 May 2026
Early-onset sepsis (EOS) — Kaiser model
Neonatal sepsis risk per 1,000 live births
wk
°C
h
GBS status
Intrapartum antibiotics
Clinical exam at evaluation
Enter GA ≥ 34 weeks and clinical inputs to estimate EOS risk.
Educational tool only — not medical advice. Simplified categorical model. For clinical use, the original Kaiser Permanente Neonatal Sepsis Calculator (Kuzniewicz 2017) at neonatalsepsiscalculator.kaiserpermanente.org is the gold standard. AAP 2018/2019 endorses risk-based approaches for late-preterm and term neonates. Reduces unnecessary antibiotic exposure substantially vs categorical algorithms (Achten 2019 JAMA Pediatr — 60-70 % reduction in antibiotic exposure with similar outcomes).
What does this mean?
Early-onset neonatal sepsis (EOS, < 72 h of life) is rare — incidence ~0.4–1 per 1,000 live births in the GBS-prophylaxis era — but devastating when missed. The trade-off: traditional categorical algorithms (CDC 2010) recommended broad empirical antibiotics for any chorioamnionitis exposure or any GBS-prophylaxis gap, leading to 10–15 % of all newborns getting antibiotics when the true sepsis rate is < 1 %. The Kaiser Permanente EOS calculator (Kuzniewicz 2017) replaced this with a Bayesian model combining maternal risk factors AND the infant’s clinical exam — and reduced antibiotic exposure by 60–70 % with equivalent outcomes (Achten 2019 JAMA Pediatr meta- analysis). Less antibiotic exposure means less microbiome disruption, better breastfeeding establishment, shorter mother-baby separation, lower NICU costs. For clinical use, the original calculator (kaiserpermanente.org) remains the gold standard.
Introduction
Early-onset neonatal sepsis (EOS) is bacterial bloodstream infection presenting in the first 72 hours of life. Incidence is ~0.5 per 1,000 live births in well-resourced settings (down from ~3 per 1,000 before universal GBS prophylaxis). Despite low incidence, mortality remains 5-15 %, motivating careful identification and prompt treatment of at-risk newborns.
The Kaiser Permanente Neonatal Sepsis Calculator (Kuzniewicz 2017) is a quantitative risk-based tool that integrates maternal intrapartum factors and newborn clinical exam to estimate EOS probability per 1,000 live births. AAP 2018/2019 endorses risk-based approaches for late-preterm and term neonates.
The model inputs
- Gestational age — baseline EOS rate varies (~0.7 per 1,000 at 34 weeks, ~0.13 at 41 weeks).
- Maternal maximum intrapartum temperature — the strongest single predictor; temperatures ≥ 38.0 °C raise risk 3-12× depending on peak.
- GBS status and intrapartum antibiotics — GBS positive without adequate antibiotics raises risk 8×; with adequate prophylaxis only 2×.
- Duration of ruptured membranes — longer ROM increases ascending infection risk.
- Newborn clinical exam — the strongest single MODIFIABLE input. Well-appearing = LR 0.41; equivocal = LR 5; clinically ill = LR 21.
The output tiers
- < 0.5/1,000 — Routine well-baby care. No additional surveillance.
- 0.5-1.0/1,000 — Enhanced observation. Vital signs q4h × 24h. Empirical antibiotics only if deterioration.
- 1.0-3.0/1,000 — Blood culture + close observation. Vital signs q2-4h × 36-48h. Empirical antibiotics considered.
- ≥ 3.0/1,000 — Empirical antibiotics. Septic workup (blood culture, FBC, CRP). Empirical IV ampicillin + gentamicin. NICU consultation.
Why risk-based outperforms categorical
Before 2015, US guidance gave empirical antibiotics for any of: maternal fever, prolonged ROM, GBS positive without adequate prophylaxis, or prematurity. This treated ~7-10 % of all newborns — approximately 400,000 US babies per year — most without sepsis. The Kaiser risk-based approach safely reduces this exposure by 60-70 % without increasing sepsis incidence (Kuzniewicz 2017; Achten 2019).
What “adequate intrapartum antibiotics” means
- Penicillin G, ampicillin, or cefazolin.
- Administered ≥ 4 hours before delivery.
- Clindamycin and vancomycin do NOT count as adequate for GBS prophylaxis.
Limitations
- Validated for ≥ 34 weeks GA. Preterm sepsis evaluation follows separate intensified pathways.
- Does not directly assess for non-GBS pathogens (E. coli, Listeria, viral) or late-onset sepsis (after 72 hours).
- Educational version uses simplified categorical LRs; the original Kaiser calculator uses logistic regression coefficients.
- Clinical judgement supersedes calculator output. A normal calculator result in a clinically ill baby still warrants empirical antibiotics; an abnormal result in a well-appearing baby on observation may not need them.
Sources
- Kuzniewicz MW, et al. A Quantitative, Risk-Based Approach to the Management of Neonatal Early-Onset Sepsis. JAMA Pediatr 2017;171:365-71.
- Puopolo KM, Lynfield R, Cummings JJ; AAP Committee on Fetus and Newborn. Management of Infants at Risk for Group B Streptococcal Disease. Pediatrics 2019;144:e20191881.
- AAP. Management of Neonates Born at ≥35 0/7 Weeks’ Gestation With Suspected or Proven Early-Onset Bacterial Sepsis. Pediatrics 2018;142:e20182894.
- Achten NB, et al. Association of Use of the Neonatal Early-Onset Sepsis Calculator With Reduction in Antibiotic Therapy and Safety: A Systematic Review and Meta-analysis. JAMA Pediatr 2019;173:1032-40.
- CDC. Prevention of Perinatal Group B Streptococcal Disease. MMWR 2010 (foundational).
Frequently asked questions
What is early-onset sepsis (EOS)?
Bacterial bloodstream infection in newborns presenting within the first 72 hours of life. EOS is most commonly caused by Group B Streptococcus (GBS) and E. coli, acquired from the maternal genital tract during labour or delivery. Incidence is ~0.5 per 1,000 live births in the US/UK (down from ~3 per 1,000 before universal GBS screening). Mortality remains 5-15 % even with prompt treatment.
What's the Kaiser EOS calculator?
A risk-based tool developed by Kuzniewicz and Escobar at Kaiser Permanente (Pediatrics 2017, JAMA Pediatr 2019). Uses maternal intrapartum factors (GA, max temp, GBS status, ROM duration, intrapartum antibiotics) plus the newborn's clinical exam to estimate probability of EOS per 1,000 live births. Stratifies newborns into routine care, enhanced observation, blood culture + observation, or empirical antibiotics tiers. Reduces unnecessary antibiotic exposure by 60-70 % compared to categorical risk-factor-based algorithms (Achten 2019 JAMA Pediatr).
Why does this matter?
Until 2015, US newborns with any of: maternal fever, prolonged ROM, GBS positive without adequate antibiotics, or prematurity, received empirical antibiotics — even if well-appearing. This led to 7-10 % of all newborns receiving 48 hours of IV antibiotics (~400,000 babies/year in the US). The risk-based approach (Kaiser calculator + 35+ week clinical exam) safely reduces this exposure substantially without increasing sepsis incidence. AAP 2018/2019 endorses risk-based approaches for late-preterm and term neonates.
When is the calculator used?
For neonates ≥ 34 weeks GA with any maternal sepsis risk factors. Calculated at birth or in the first 1-2 hours of life. NOT validated for < 34 weeks GA — preterm sepsis evaluation follows separate intensified pathways. Performed before deciding whether to start empirical IV antibiotics, send blood culture, or observe.
What does 'adequate intrapartum antibiotics' mean?
Per CDC 2010 and AAP 2019: penicillin G, ampicillin, or cefazolin administered ≥ 4 hours before delivery. Clindamycin and vancomycin are NOT considered adequate for GBS prophylaxis. Antibiotic timing matters — even penicillin given 2 hours before delivery is incomplete prophylaxis.
What if the clinical exam is equivocal?
About 5 % of newborns have transient signs (mild tachypnoea, irritability, tone changes, transient need for FiO2 > 21 %) without true sepsis. The Kaiser model weights equivocal clinical exam with a likelihood ratio of ~5. Combined with maternal risk factors, this often triggers blood culture and enhanced observation rather than immediate empirical antibiotics. Re-exam at 1-2 hour intervals; many equivocal findings resolve.
What if blood culture comes back positive?
Full course of IV antibiotics. Most positive cultures are GBS or E. coli — both sensitive to ampicillin + gentamicin (the standard empirical regimen). Duration: 7-10 days for confirmed sepsis without meningitis; 14-21 days if meningitis is confirmed. NICU-level care for the duration. Follow-up audiology, neuroimaging, and developmental review.